The problem of characterizing the manifest probabilities of a latent trait model is considered. The approach taken here differs from the standard approach in that a population of examinees is being considered as opposed to a single examinee. Particular attention is given to the Rasch model. (Author/JKS)

The magnitude of statistical bias for the phi-coefficient was investigated, using computer simulated examinations in which all the students had equal knowledge. Several modifications of phi were tested, but when applied to real examinations, none succeeded in improving its reproducibility when items are re-used on equivalent student groups.…

A modification of the method of reciprocal averages for scaling multiple choice data is proposed. The proposed method handles the data in a piecewise fashion and allows for faster convergence to a solution. (Author/JKS)

Brennan's B statistic is a generalized upper-lower discrimination index which was first published in 1972: Peirce earlier introduced a statistic on the relation between a predictor and an outcome which has the same structure as Brennan's B. (Author/CP)

An alternative derivation was given of Gaylord's formulas showing the relationships among the average item intercorrelation, the average item-test correlation, and test reliability. Certain parallels were also noted in analysis of variance and principal component analysis. (Author)

A computer-assisted examination program, written in TUTOR for PLATO IV at the School of Basic Medical Sciences, University of Illinois College of Medicine, is discussed. The system is automated from the scheduling and delivery of basic medical examinations to the analysis of student performance data. (RAO)

Order analysis, a technique to isolate unidimensional hierarchies representing multidimensional structure of binary data, is reviewed. Several theoretical flaws inherent in the probalistic version are presented. Suggestions of possible directions for future research are offered. (Author)

A new measure of similarity between persons applicable in Q-analysis is proposed. It allows assumptions of non-orthogonality between the items, across which the similarity is computed. The similarity measure may also be applied in an R-analysis. (Author/JKS)

The Test Analysis Program (TAP) is a comprehensive, flexible computer system designed to score and to analyze objective educational tests. The goals of the designers were to construct a program which would be user-oriented, flexible, and clear in structure and in output. (Author/JKS)

Items designed to measure the ability to derive the meanings of words from context were drawn from seven published reading tests. The items were administered to high school students without the context. Only one published test had vocabulary-in-context items that were consistently context-dependent. (Author)

Procedures for the Rasch model, sample free item calibration are reviewed and compared for accuracy. The theoretically ideal procedure is shown to have practical limitations. Two alternatives to the ideal are presented and discussed. A correction for bias in the most widely used alternative is presented. (Author/JKS)

A similarity-based smoothing approach to nondimensional item analysis was studied. Simulated and actual data are presented to show that when responses are determined by a latent ability variable, this similarity-based smoothing procedure can reveal the dimensionality of ability satisfactorily. (SLD)

Traditional Item Selection Rules (TISRs) were compared to Intervention Item Selection Rules (IISRs) on the same set of alcohol attitude items from archival data, producing scales with differing psychometric properties. A cross-validation study was run. Guidelines for selecting change-sensitive items, problems and advantages of IISRs are…

Presents the hierarchical generalized linear model (HGLM) as an explicit two-level formulation of a multilevel item response model. Shows that the HGLM is equivalent to the Rasch model, and that a characteristic of the HGLM is that person ability can be expressed as a latent regression model with person-characteristic variables. Shows that the…

Discusses the logic of the rule-space model (K. Tatsuoka, 1983) as it applies to test development and analysis. The rule-space model is a statistical method for classifying examinees' test item responses into a set of attribute-mastery patterns associated with different cognitive skills. Directs readers to a tutorial that may be downloaded. (SLD)